How the numbers are built

Fees vs revenue

Fees are what users paid. Most of it goes to liquidity providers, creators or other third parties. Revenue is what the protocol and its token holders keep, and it has three parts: treasury income, payouts to holders, and buybacks or burns funded from fees. That is the equity convention: earnings count before the company decides between dividends and buybacks. A buyback funded from the treasury is a use of revenue already counted, so it is not added twice. Everything else is excluded: LP fees, creator shares, fees that buy other tokens, token sales, and burns of pre-held tokens.

Holder revenue and P/HR

Holder revenue is the slice that reaches holders directly: payouts, buybacks and burns. P/HR = market cap ÷ (30-day holder revenue × 365 ÷ 30). A protocol can have healthy revenue and zero holder revenue when everything stays in a treasury. Each stream is tagged code when a contract routes it without a human decision, or discretionary when a key can start, stop or redirect it. A discretionary buyback counts when it happens and deserves a lower multiple than an enforced one, because it can stop tomorrow.

Choosing the window

The default is the last 7 days annualised, and every headline multiple on the site uses it. Seven days is the shortest window with enough days behind it to survive one loud trading session, and on a chain where most protocols are a few weeks old it tracks the current run rate far better than 30 days does. A 30-day window here still carries weeks that no longer describe the protocol, and for anything that started earning three weeks ago it divides by days that never existed.

The ranking and the chart carry a window switch: 24h, 7d, 30d, or since launch. Every window is annualised so the multiples stay comparable. 24h multiplies the last day by 365, 7d multiplies the last week by 52.1, 30d multiplies the last month by 12.2, and since launch divides everything the protocol has ever earned by the days since it first earned anything, then multiplies by 365.

The window matters more than it looks on a chain this young. A protocol that started routing fees three weeks ago still gets divided by 30 in the 30-day column, which understates its run rate and pushes its multiple up. A protocol having one enormous day looks cheap on the 24h column and will not stay there. A dagger marks any row where the protocol has been earning for less time than the selected window, and the since-launch column shows the day count it is dividing by. Read at least two windows before believing either.

The trend column

A trailing window hides decay. StonkBrokers sat at the top of this table as the cheapest protocol on the board while its revenue was already down about 75%, because the 30-day window still held the weeks before the fall. Shroom did the same on the way down in September: four strong days made a week look healthy after daily payouts had dropped 94%.

Trend uses whichever of two comparisons the protocol has the history for. Where it has run for at least two of the selected windows, it compares the window with the window immediately before it. Where it has not, the figure carries an asterisk and compares the last 24 hours with the window’s own daily average instead. The second basis exists because the first is useless on a young protocol: Shroom’s prior seven days held $956 against $504,000 in the current seven, which reads as plus fifty-two thousand percent and tells you nothing. Negative means the protocol is earning less now than the multiple beside it assumes.

Price to revenue

P/R = market cap ÷ (30-day revenue × 365 ÷ 30). The FDV version swaps in fully diluted value. Both are shown because low-float tokens look cheap on one and expensive on the other. UP is the clearest case: about 3% of supply floats.

Sources

Launch-bundle screen

Replay the token from its mint. Find the first pool. Decode every swap in the creation block plus 50 blocks (about 5 seconds on this chain). Group buys by same block, same gas price and same amount. Trace each origin's first token funding. Report the share of supply the cohort took and how much has since been sold.

Verdicts: CLEAN means no coordination signal in the window. COORDINATED means same-block, same-gas, same-amount groups or shared funders took a material share. BUNDLED means a pre-funded fleet took most of the supply in seconds. Timing and gas never prove common ownership; they prove coordination of behaviour.

Token audit (ours)

Eight surfaces, each rated from on-chain evidence pinned at a block: A privileged controls, B liquidity custody, C sellability at size (Quoter ladders plus executed sells), D concentration (full Transfer replay), E launch, F fees and where they go, G holder rights, H external dependencies. A critical on any surface caps the verdict. GO-WITH-CONDITIONS means no executable path was found for a privileged actor to mint, freeze or pull principal, with named conditions. NO-GO means a stated requirement fails. Never read either as safe.

Professional audit

Separate from our token audit. We check whether a recognised security firm (Trail of Bits, OpenZeppelin, Spearbit, Cyfrin, Zellic, Halborn, CertiK, Sherlock, Certora and peers) has published an audit of the protocol's contracts: the project's site and docs, its GitHub, the DefiLlama audit fields, and the firms' own report archives. Audited means we found the report or a firm-named listing. Self-reported means the project claims an audit but names no firm or publishes no report. No audit found means a real search turned up nothing. We also note whether the audited scope matches the contracts deployed on Robinhood Chain: an audit of Uniswap v4 core covers the same bytecode here; an audit of a launchpad template on another chain may not cover a fork. A clean professional audit says the code does what it says. It says nothing about who holds the keys, how the launch went, or whether holders have any rights. That is what the token audit is for.

Refresh

DefiLlama and DexScreener are re-pulled every hour by a scheduled job. The JSON behind the site is public at /api/data. Curated fields (links, contracts, screens, audits) live in a versioned file and change when a new run lands.

← back to the ranking